Development of breast cancer risk model based on estrogen metabolomics. Anti-estrogens provide an effective strategy to prevent breast cancer but eligible women generally decline therapy because of unfavorable benefit/risk ratios. Data from prevention studies indicate that fifty women need to be treated with these agents for five years to prevent one breast cancer. To improve the ratio of benefit to risk, a more powerful method of identifying women at very high risk of developing breast cancer is urgently needed. Prediction of disease risk is best grounded on factors involved in its pathogenesis. We propose an integrative hypothesis regarding the carcinogenic process which involves both estrogen receptor alpha (ER) dependent as well as ER independent actions. Through ER, estradiol (E2) stimulates proliferation with resultant replicative mutations and promotes the growth of cells harboring those mutations. Independent of ER, estrogen metabolites both form unstable DNA adducts and generate oxygen free radicals thorough redox cycling to initiate mutations. Several genetically regulated enzymes modulate estrogen metabolism and the process of repair of estrogen induced mutations. Our innovative hypothesis regarding estrogen induced breast cancer integrates all of these processes into a model of carcinogenesis and implicates the entire estrogen metabolome in the genesis of breast cancer. These concepts suggest that measurement of estrogen metabolomics should provide a powerful, mechanism-based method of predicting who will develop breast cancer. Metabolomic assessment entails quantitative measurement of aromatizable androgens, estrogens, and their metabolites and SNPs from enzymes regulating the metabolic process. Several important factors have recently come together to enable us to test this concept. A new, state of the art, mass spectrometer coupled with an ultra-performance liquid chromatography system makes it possible for the first time to measure all estrogen metabolites in small amounts of serum. We can measure SNPs which involve enzymes regulating estrogen metabolism and have been shown to correlate with breast cancer risk. To develop a model, we will utilize serum samples and risk factor data from 3 cohort studies (NYU, Clue I and II, and Rancho Bernardo) which had collected blood from women 5-20 years ago and then followed them prospectively for development of breast cancer. Availability of these techniques, samples and risk factor data allows performance of a nested case-control study to develop a new, more powerful risk prediction model. We will then validate this model in a completely independent data set involving the French Teacher's Study. We anticipate reducing the number of women needed to be treated to prevent one breast cancer from 50 to 13 with tamoxifen and to 5.with aromatase inhibitors.